This article is more than 1 year old
Darth Vader voice actor James Earl Jones allows AI to take over the role
Ukrainian outfit Respeecher's GAN-based neural audio enhancer embraces The Dark Side
James Earl Jones, the actor who has voiced iconic Star Wars villain Darth Vader since 1977, has reportedly permitted his past utterances to be fed into an AI that will ensure his distinct tones become replicable once he becomes one with the Force.
News that Vader will achieve digital immortality comes from report in Vanity Fair which reveals that a Ukrainian company called Respeecher was hired by Lucasfilm to reproduce Jones's famous baritone for the Obi-Wan Kenobi miniseries recently released on the Disney+ streaming service.
Respeecher describes itself as offering content creators to "Create speech that's indistinguishable from the original speaker."
That mattered to the producers of Obi-Wan Kenobi because the show was set between the events depicted in Star Wars Episodes III and IV, and they wanted Vader's voice to reflect that of a younger villain instead of Jones's rather more mature voice.
The story also states that Jones has allowed his archival material to be absorbed by Respeecher for later re-use, as at age 91 Jones has expressed a desire to step back from the character.
Your correspondent thinks Resepeecher could also do worse than to secure the rights for Jones's sublime turn as Thulsa Doom in 1982's Conan the Barbarian, in which Arnold Schwarzenegger brought himself to Hollywood's attention after grunting his way through the titular role.
But we digress.
- Pretend starship captain to take trip in real space capsule
- Pentagon scraps $10bn JEDI winner-takes-all cloud contract
- Luke Skywalker used to bullseye womp rats in his T-16 on Tatooine. But Star Wars: Squadrons misses the mark
Respeecher continues to operate from Ukraine. The company describes its wares as a version of the "super resolution" technology used to upscale images captured by magnetic resonance imaging.
"An artificial neural network adds missing details in the time domain by increasing the effective sampling rate of an audio signal," explains a white paper that details the company's tech.
"In the nutshell, our super resolution network is a GAN-based neural audio enhancer that adds extra resolution to recordings with limited bandwidth," the document adds. "The enhancement is performed by an artificial neural network that analyzes the frequency range of the input low resolution audio and completes its spectrum by generating a high frequency signal that blends smoothly with the original audio."
It seems apt to end with a Darth Vader quote, spoken in a conference room on the Death Star in Episode IV.
"Don't be too proud of this technological terror you've constructed. The ability to destroy a planet [or digitally clone a voice] is insignificant next to the power of the Force."
And you totally read that in James Earl Jones's voice. ®